Investigation of pilot training effect in massive-MIMO TDD system

Massive-MIMO system is widely addressed by the literature as one of the main contributors for the next generation of the cellular network (5G). In the Massive-MIMO system the number of transmit antennas is much larger than the number of served users. Hence, it promises higher spectral efficiency by improving the rank of the overall channel matrix and localizing the interference. To mitigate interference (especially between adjacent sub-sectors), the accuracy of estimated channel is highly important. While, channel estimation in Massive-MIMO systems will cause the technical challenge of pilot contamination, due to the large number of antennas. In this regard, this paper addresses the effect of pilot training during channel estimation in downlink and uplink transmission of TDD Massive-MIMO system. The results present the performance of channel estimation in term of mean square error under different channel estimation techniques, various number of transmit antennas, and different length of pilot training. Moreover, results are compared under two different scenarios; without and with interference (pilot contamination). The results highlight the importance of pilot contamination avoidance in the Massive-MIMO system.